Matched set of tubes, bias questions

[pet0r]

Yesterday I installed a new matched quad of Shuguang EL-34B tubes in my DR103. Two things I noticed:

1 - they do not seem to be drawing the same current, two of them around 21ma, the others around 30.
2 - according to these values, they are biased quite low with the original bias cicruit. Voltage is around 475V. I will need LESS neg voltage! Is that normal when most people need voltage doublers?

Now recently I've bought the same Chinese tubes from this seller and put them in my other DR103. They were perfectly matched, so I trust the seller knowing his stuff.

The previous set of tubes in the amp I retubed yesterday was also a little off, two tubes drawing more current than the others. And now again. So I'm thinking the 'mismatch' is caused by something my amp.
Any idea which components in my amp I could check to see if anything's off? The 100k bias resistors? The 22k grid stoppers? An unbalanced PI?

Thanks for helping me out!
Peter

[pet0r]

By the way, on this amp I already replaced

- all four power tube sockets with nos McMurdo's, old ones were in pretty bad shape
- both the 47nF PI output caps with nos mustards

and checked the 100K bias resistors, which measure ok.

[mhuss]

Is it the two left ones and two right ones (versus random)? Does the mismatch follow the tubes when you swap them left to right? As a sanity check, measure the voltage on pin 5 of each of the output tubes, they should all be the same.

Also, did it have this problem before you replaced the PI coupling caps?

--mark

[pet0r]

I'm not at home right now, but if I remember correctly, it's the outer 2 tubes that measure low, while the middle ones measure higher. Certainly each pair of 2 has one tube that measures lower than the other.
I tried swapping the tubes around, but as I remember that did not make a difference. I'll check to make sure tonight.

What I do know for certain is that the problem was there before I changed the PI caps.

[pet0r]

Hmm it might not be as bad as I thought.

The highest one measures 26mA, the lowest 21.3mA. I switched one pair around, and measurements follow the tubes. So it's probably the tubes themselves. Voltage on pin 5 is the same on all four.

[pet0r]

Would this still qualify as a matched set of tubes?

[Dr.HI-TONE]

unfortunately by most resellers

[mhuss]

Yes, "matched " is a relative term, sigh.

--mark

[pet0r]

If the 47n PI caps are good, the 2x 100k bias and 4x 22k grid resistors are good, and the bias filtercap is good, can there be anything else faulty in the amp?
Perhaps the 82k/91k PI resistors or the transformer?

Or is it really just down to the tubes?

[mylescdavis]

I've read that, in a four-tube output stage, the most important thing is whether each *side* is balanced with respect to the other - in other words, add the two current draws for one side, compare with the total for the other two. In this case, it sounds like two tubes are lower than the other two, so you make sure that you have one low and one high on each side.

Or am I talking crazy talk?

[mhuss]

It's a toss-up. It you have one high and one low on each side, the push and pull are better matched, but you still have two tubes carrying more than their fair share of the load.

--mark